Class A Scavenger Receptor Up-regulation in Smooth Muscle Cells by Oxidized Low Density Lipoprotein*

Oxidative stress caused by phorbol esters or reactive oxygen up-regulates the class A scavenger receptor (SR-A) in human smooth muscle cells (SMC), which normally do not express this receptor. The increase in SR-A expression correlates with activation of the redox-sensitive transcription factors activating protein-1 c-Jun and CCAAT enhancer-binding protein β. Here we show that coincubation of SMC with macrophages or oxidized low density lipoproteins (LDL) from macrophage-conditioned medium activates these same regulatory pathways and stimulates SR-A expression. The increased SR-A gene transcription induced by cell-oxidized LDL up-regulated SR-A mRNA and increased by 30-fold the uptake of acetyl LDL, a ligand for the SR-A. Copper-oxidized LDL also increased SR-A receptor expression. Oxidized LDL with a lipid peroxide level of 80–100 nmol/mg of LDL protein and an electrophoretic mobility ∼1.5 times that of native LDL exhibited the greatest bioactivity. Inhibition of calcium flux suppressed SR-A induction by oxidized LDL. Conversely, calcium ionophore greatly enhanced SR-A up-regulation by oxidized LDL or other treatments that promote intracellular oxidative stress. This enhancement was dependent upon concurrent up-regulation of SMC cyclooxygenase-2 expression and activity and was blocked by the cyclooxygenase-2 inhibitors NS-398 and Resveratrol. In THP-1 cells, oxidized LDL induced monocyte-to-macrophage differentiation and increased SR-A expression. These findings support a role for mildly oxidized LDL in the redox regulation of macrophage differentiation and SR-A expression and suggest that increased vascular oxidative stress may contribute to the formation of both SMC and macrophage foam cells.

The traditional risk factors for atherosclerosis, which include hypercholesterolemia, hypertension, cigarette smoking, diabetes, and high fat diet, have all been associated with endothelial dysfunction (1). Under these conditions, circulating monocytes adhere to the arterial endothelium, migrate to the subendothelial space, and differentiate into resident macrophages within the subendothelial cell matrix. The differentiated cells express scavenger receptors that take up modified lipoproteins, leading to a massive accumulation of cholesterol esters and the appearance of foam cells. These macrophage-derived foam cells make up the fatty streak lesions that precede more advanced atherosclerotic lesions, which can ultimately cause thrombosis and myocardial infarction (2).
Advanced atherosclerotic lesions also contain smooth muscle cells (SMC), 1 which migrate from the media of the blood vessels to the neointima, where they proliferate. Intimal SMC can also accumulate large amounts of cholesterol esters and become foam cells. Although the cholesterol that accumulates in foam cells is derived from lipoproteins, the mechanisms by which this occurs are not fully understood. Plasma lipoproteins are oxidatively modified in the subendothelial matrix, where they cause the release of cytokines that attract monocytes to the subendothelial space (3). With further oxidation, these low density lipoproteins (LDL) become ligands for several scavenger receptors on macrophages. Since they are not downregulated as the intracellular content of cholesterol increases, these receptors are thought to contribute to the excessive uptake of modified lipoproteins and to the lipid engorgement characteristic of macrophage-derived foam cells (4). The first of these scavenger receptors to be cloned was the class A scavenger receptor (SR-A) (5). Knockout mice lacking this receptor were resistant to the development of atherosclerosis, suggesting that the SR-A contributes to the uptake of modified lipoproteins and to cholesterol ester accumulation in macrophages in vivo (6).
How SMC in the artery wall accumulate lipid is less clear, since SMC were initially thought to be devoid of scavenger receptors. We have shown, however, that SMC express SR-A and that receptor expression is regulated over a wide range (7)(8)(9)(10). SR-A activity is induced in SMC by treatments that increase intracellular oxidative stress, such as phorbol esters and the combination of H 2 O 2 and vanadate (11). Certain growth factors that increase SMC SR-A activity (10) have also been associated with increased oxidative stress, including interleukin-1, tumor necrosis factor ␣ (12), epidermal growth factor (13), platelet-derived growth factor (10,14), and transforming growth factor ␤ (10, 15). The latter two factors are responsible for the up-regulation of SMC SR-A activity by platelet secretory products (10). Reactive oxygen species (ROS), such as superoxide anions and H 2 O 2 , function as second messengers in signal transduction, mediating ligand stimulation by tyrosine phosphorylation (16). Ca 2ϩ signaling is also thought to contribute to ROS-induced gene expression (17,18). Downstream targets of ROS include the stress-activated c-Jun amino-terminal kinase (JNK) (19), which phosphorylates and activates c-Jun. Indeed, the level of intracellular glutathione is a key regulator of JNK induction and therefore of activating protein-1 (AP-1)/c-Jun transcriptional activity (20). A composite AP-1/Ets binding element located between Ϫ67 and Ϫ50 base pairs (bp) relative to the transcriptional start site of the SR-A gene is critical for regulation of macrophage SR-A activity (21). We previously showed that the increase in SMC SR-A expression under conditions that promote oxidative stress correlates with the activation of the AP-1/c-Jun transcription factor and increased JNK activity (22). Binding of CCAAT/enhancer-binding protein (C/EBP) ␤ to a C/EBP site in the SR-A promoter between Ϫ44 and Ϫ21 (bp) relative to the transcriptional start site was also necessary for full up-regulation of SR-A expression in SMC (22).
SMC SR-A activity is up-regulated in vivo by an atherogenic diet in rabbits (23), suggesting that hyperlipidemia also contributes to intracellular oxidative stress, thereby altering SMC gene expression either directly or through paracrine factors secreted by local inflammatory cells. Cytokines known to be secreted by activated leukocytes up-regulated SR-A expression in rabbit SMC in vitro (23). In developing atherosclerotic lesions, intimal SMC are closely associated with macrophages that secrete numerous growth factors, cytokines, and ROS (24), suggesting that factors secreted by macrophages might interact with SMC to increase SR-A expression. In this study, we tested the hypothesis that coculture of SMC with macrophages would increase the SMC SR-A activity. Here we report that coculture with macrophages increased SR-A expression in SMC; however, the enhanced SR-A activity was caused by cell-oxidized LDL and not by growth factors or cytokines secreted by the macrophages. Further characterization of the mechanism of receptor up-regulation revealed a role for calcium flux and cyclooxygenase (COX)-2.

EXPERIMENTAL PROCEDURES
Materials and Cells-Fetal bovine serum (FBS), Dulbecco's phosphate-buffered saline, and the fluorescent probe 1,1Ј-dioctadecyl-3,3,3Ј,3Ј-tetramethylindocarbocyanine perchlorate (DiI) were obtained as described previously (7). Phorbol 12-myristate 13-acetate, sodium orthovanadate (vanadate), H 2 O 2 , Resveratrol, EGTA, BAPTA, diltiazam, and verapamil were from Sigma. Mibefradil was obtained from Hoffman-La Roche (Basel, Switzerland). The COX-2 inhibitor NS-398 (25) was obtained from Biomol Research Laboratories (Plymouth Meeting, PA). The concentration of active H 2 O 2 was standardized immediately before each use by spectrophotometry on the basis of an extinction coefficient of 0.0393 at 240 nm. Primary human aortic SMC and SMC growth medium were purchased from Clonetics (San Diego, CA). SMC lines were maintained at 37°C in 7% CO 2 . THP-1 cells (American Type Culture Collection, Manassas, VA) were maintained in RPMI (Life Technologies, Inc.) at 37°C in 5% CO 2 . Ham's F-10 medium was obtained from Life Technologies. Rabbit peritoneal macrophages were obtained from New Zealand White rabbits by peritoneal lavage 3 days after stimulation with an intraperitoneal injection of mineral oil (26). The macrophages were isolated and washed in phosphate-buffered saline, resuspended in RPMI, and added immediately to confluent SMC cultures.
Lipoproteins and Conditioned Media-Human LDL (d ϭ 1.02-1.05 g/ml) were obtained from plasma and were either labeled with DiI and acetylated (Ac) as described (9,27) or oxidized. For cell-mediated oxidation of LDL, LDL (100 g/ml) in serum-free Ham's F-10 medium were incubated with differentiated THP-1 cells at 37°C for 6 h, unless otherwise specified. LDL were also oxidized in vitro with copper sulfate (10 M) for 6 h at 37°C. Oxidation was stopped by adding EDTA and butylated hydroxytoluene to final concentrations of 2 mM and 40 M, respectively. The cell-oxidized LDL were dialyzed against phosphatebuffered saline containing 100 M EDTA and stored at 4°C for no more than 3 days before use. The extent of LDL modification was assessed by electrophoretic mobility in 1% agarose. Lipid peroxide levels in oxidized LDL were measured by quantifying the oxidation of iodide to iodine (I 3 ), as described (28). Thiobarbituric acid-reactive substances in preparations of cell-modified LDL were measured as described (29).
To prepare lipid-poor serum, human plasma was incubated with Cabosil (25 mg/ml) (Packard International, Zurich, Switzerland) for 2 h at 37°C, 15 h at 4°C, and then centrifuged at 100,000 ϫ g. The lipid-depleted supernatant was decanted and dialyzed against phosphate-buffered saline. To prepare cell-conditioned medium, THP-1 cells were differentiated for 72 h with phorbol 12-myristate 13-acetate (100 nM), washed extensively, and incubated for 6 h in RPMI with or without 10% FBS. Conditioned medium was cleared of cellular debris by low speed centrifugation followed by passage through a 0.45-m filter, aliquoted, and stored at Ϫ70°C. Activity was not compromised by freezing and thawing.
To determine if the bioactivity of conditioned medium was associated with the cell-modified LDL, the LDL were reisolated from the medium by fast performance liquid chromatography (FPLC) or ultracentrifugation as described (30). The FPLC fractions were tested for their ability to up-regulate SMC SR-A activity, and the bioactivity was compared with the LDL cholesterol elution pattern. The top and bottom thirds of the centrifuge tube, containing LDL and no LDL, respectively, were similarly tested.
Electrophoretic Mobility Shift and JNK Assays-Nuclear extracts from human SMC treated with cell-modified LDL or with 100 M H 2 O 2 and 10 M vanadate (H/V) were prepared as described (22). For binding studies, nuclear proteins (5 g) were incubated with the AP-1/Ets site at Ϫ67 to Ϫ50 bp in the SR-A promoter or with the C/EBP consensus wild-type and mutant binding elements (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) as described (22). Antibodies to c-Jun, c-Fos, C/EBP␤, and GAD153 used for supershift analyses were from Santa Cruz Biotechnology. Total cell extracts were prepared from similarly treated human SMC for the JNK assays as described (22) and assayed for the ability to phosphorylate a glutathione S-transferase-c-Jun fusion protein bound to glutathione-agarose beads.
RNA Analyses-Total RNA was isolated by phenol/chloroform extraction with Trizol (Life Technologies) and quantified spectrophotometrically by absorption at 260 and 280 nm. mRNA was purified with Dynabeads oligo(dT) (Dynal, Lake Success, NY) and magnetic separation techniques. Full-length first-strand cDNA was generated with avian myoblastosis virus reverse transcriptase, a cDNA cycle kit (Invitrogen, Carlsbad, CA), and random primers; human SMC mRNA (50 ng) or total THP-1 RNA (5 g) was used as a template. Subsequent amplifications by polymerase chain reaction (PCR) (30 cycles) were performed with constant amounts of the cDNA template, 1 unit of Taq polymerase, and specific sets of primers (0.1 M) for the COX-2, type I and II SR-A, CD36, and glyceraldehyde-3-phosphate dehydrogenase control genes. The sequences for the primers used to amplify human COX-2 mRNA were kindly provided by Dr. Andrew Dannenberg (Cornell Medical Center, New York, NY): sense COX-2 (5Ј-GGTCTGGTGC-CTGGTCTGATGATG) and antisense COX-2 (5Ј-GTCCTTTCAAGGAG-AATGGTGC) (PCR product 724 bp). The sense primer for the SR-AI and SR-AII genes was selected from a sequence common to the two subtypes: 5Ј-GATTGGGAACATTCTCAGACCTT. For the SR-AI and SR-AII genes, the antisense primers were 5Ј-CTTGTCCAAAGTGAGCTGC-CTT (PCR product 444 bp) and 5Ј-CTGCCCTAATATGAT CAGTGAGT (PCR product 288 bp), respectively. For the CD36 gene, the sense primer was 5Ј-ATGGGCTGTGACCGGAACT, and the antisense primer was 5Ј-ACAGACCAACTGTGGTAG (PCR product 604 bp). A control primer set for the human glyceraldehyde-3-phosphate dehydrogenase gene was from CLONTECH (PCR product 983 bp).

Regulation of Human SMC SR-A by Macrophages-Platelet
secretory products and ROS, as well as certain cytokines and growth factors alone and in combination, synergistically increase SR-A expression in SMC (7, 10, 22, 23, 31). Since activated macrophages in atherosclerotic lesions secrete several of these factors and are in intimate contact with SMC, we hypoth-esized that macrophage products might up-regulate SMC SR-A activity. To test this hypothesis, we cocultured SMC and rabbit peritoneal macrophages for 24 h and used fluorescence microscopy to assess the uptake of DiI-labeled AcLDL. Coincubated SMC internalized labeled AcLDL, indicating increased SR-A activity, whereas the control cells did not (Fig. 1). Conditioned medium alone also increased SR-A activity, indicating that cell-cell contact was not required (Fig. 2). However, the increased activity did require lipoproteins. Conditioned medium containing LDL (100 g/ml) or 10% FBS increased SMC SR-A activity 25-30-fold, whereas lipoprotein-or serum-free conditioned medium was inactive (Fig. 2).
Next, we determined whether the bioactivity resided within the LDL or was mediated by factors secreted by the macrophages in response to incubation with LDL. LDL were isolated from conditioned medium by FPLC and tested for their ability to up-regulate SMC SR-A activity. The column fractions that contained the LDL, as determined by cholesterol content, coincided with the fractions that up-regulated SMC SR-A activity (Fig. 3). To confirm that the LDL fraction contained the active component, LDL (d ϭ 1.063 g/ml) were reisolated from conditioned medium by ultracentrifugation and tested for bioactivity. The uptake of DiI-labeled AcLDL increased markedly in SMC incubated with the "top" lipoprotein-containing fraction (21.8 Ϯ 0.9-fold over control cells); the "bottom" nonlipoprotein fraction had minimal activity (2.7 Ϯ 0.3-fold increase over control).
To rule out the possibility that cytokines or growth factors secreted by the macrophages contributed to the increased SMC SR-A activity, we performed antibody blocking experiments. Under conditions that prevent platelet secretory products from increasing SR-A activity in SMC (10), antibodies to interleukin-1, interleukin-6, platelet-derived growth factor, fibroblast growth factor, insulin-like growth factor, and transforming growth factor ␤, alone and in combination, did not block the bioactivity in the conditioned medium (data not shown).
Chemically Oxidized LDL Up-regulate SMC SR-A Activity-Because LDL incubated with cells undergo oxidative modification, we hypothesized that oxidatively modified LDL were the active species in the conditioned medium. We tested this hypothesis in three ways. First, the active preparations of LDL from cell-conditioned medium were assayed for lipid peroxide level and electrophoretic mobility. The maximal ability to increase SR-A activity correlated with mild lipoprotein oxidation, as indicated by a peroxide level of 80 -100 nmol/mg LDL protein and a relative electrophoretic mobility ϳ1.5 times greater than that of native LDL (data not shown). The thiobarbituric acid-reactive substance content of the active LDL preparations was consistently Ͻ10 nmol/mg LDL protein. More extensively oxidized LDL were toxic to the SMC (data not shown).
Second, LDL were incubated for 6 h with THP-1 cells in the presence of EDTA (200 M) and butylated hydroxytoluene (20 M) to prevent oxidation. Under these conditions, no lipid peroxides formed, the electrophoretic mobility of the LDL was unchanged, and the conditioned medium did not up-regulate SR-A (data not shown).
Third, we evaluated the ability of mildly chemically oxidized FIG. 2. Effect of medium conditioned by incubation with THP-1 cells for 6 h on SMC SR-A activity. SMC were incubated for 15-18 h in conditioned medium with or without LDL (100 g/ml) or 10% FBS, washed, and incubated for 15-18 h in fresh medium containing DiI-labeled AcLDL (5 g/ml). The cells were then trypsinized, fixed, and assayed for SR-A activity by fluorescence-activated cell sorter analysis. Data are reported as the increase in the mean fluorescence intensity of treated cells relative to that of control cells. Results represent the mean Ϯ S.D. of three independent experiments. Conditioned medium containing LDL or FBS increased the uptake of DiI-labeled AcLDL by SMC, indicating increased SR-A activity. Conditioned medium lacking serum and lipid-poor conditioned medium were inactive.

FIG. 1. Effect of coincubation with macrophages on SMC SR-A activity.
Human SMC were grown either alone (left panels) or together with freshly isolated rabbit peritoneal macrophages (50:1, SMC:macrophages) (right panels) in medium containing 10% FBS. After 15 h, the cells were washed vigorously, incubated for 24 h in medium containing DiI-labeled AcLDL (5 g/ml), and examined by phase-contrast (upper panels) and fluorescence microscopy (lower panels). Only SMC that had been coincubated with macrophages incorporated the labeled AcLDL (lower right).
LDL to activate SMC SR-A expression. Chemically modified LDL with lipid hydroperoxide levels and electrophoretic mobilities similar to those of cell-modified LDL up-regulated SMC SR-A activity 10-fold more than native LDL; however, the level of SR-A activation was notably less than when cell-modified LDL were used (Fig. 4). This difference in bioactivity might have resulted from differences in the specific oxidized lipids formed by the two oxidation methods or from effects of additional bioactive factors released by activated macrophages into the medium.
AP-1/c-Jun Is Induced and Activated in SMC by Cell-modified LDL-Up-regulation of SR-A activity in SMC treated with phorbol ester or H/V is dependent upon increased binding of the AP-1/c-Jun transcription factor to SR-A promoter elements (22). We therefore hypothesized that incubation of SMC with cell-modified LDL would increase the level of this transcription factor. Electrophoretic mobility shift assays showed a time-dependent increase in AP-1 binding in nuclear extracts from human SMC incubated with cell-modified LDL; supershift analyses identified c-Jun and c-Fos in the binding complex, and competition studies confirmed the specificity of binding (Fig.  5A). Because AP-1/c-Jun must be phosphorylated to be active, we examined the activity of the specific c-Jun-activating kinase, JNK. Increased JNK activity was present within the SMC lysates after incubation with conditioned medium (Fig. 5B).
C/EBP␤ Is Induced and Processed in SMC Treated with Cell-modified LDL-The SR-A promoter also contains a C/EBP binding site, and full SR-A transcriptional activity in SMC requires C/EBP␤ as well as AP-1/c-Jun binding (22). Electrophoretic mobility shift assays showed greater C/EBP binding to nuclear extracts from SMC treated with cell-modified LDL than to nuclear extracts from untreated control SMC (Fig. 6A). The binding activity was supershifted with C/EBP␤ antibody but not with antibody to the related transcription factor GADD 153 (Fig. 6B). The specificity of C/EBP␤ binding was further demonstrated by complete competition with unlabeled oligonucleotide corresponding to the consensus C/EBP wild-type sequence but not with a mutant sequence containing an 8-bp substitution in the binding motif. The central band in the tripartite binding complex was consistently more prominent at the later time points (Fig. 6B). Variable band intensities in the complex C/EBP␤ binding pattern are due to the differential translational and posttranslational processing characteristic of the C/EBP␤ trans-activator protein (32).
Cell-modified LDL and Calcium Ionophore Increase SR-A Activity-Treatment with H/V increases intracellular oxidative stress and activates redox-sensitive SR-A gene expression in SMC (22). Because oxidants stimulate calcium signaling (17,18), we examined the effect of the calcium ionophore A23,187 on the ability of H/V or cell-modified LDL to regulate receptor activity (Fig. 7A). Incubation of SMC with A23,187 and either H/V or cell-modified LDL significantly increased SR-A activity. The up-regulation was 3.4-fold greater than was obtained with H/V alone and 2.2-fold greater than with conditioned medium alone. Cotreatment with nickel chloride, a calcium channel blocker that also exerts intracellular effects on calcium, completely blocked the ionophore-induced increase in SR-A activity The fractions that induced SR-A activity in SMC coincided with the fractions that contain LDL cholesterol. (Fig. 7B). Three specific calcium channel blockers, diltiazam, verapamil, and mibefradil, did not suppress SR-A up-regulation by either H/V or cell-modified LDL (data not shown). However, chelation of either extracellular or intracellular calcium with EGTA or BAPTA, respectively, partially blocked the up-regulation of SR-A activity by H/V plus A23,187 (Fig. 7B). Together, they almost completely inhibited SMC SR-A up-regulation by H/V plus A23,187. SR-A up-regulation by cell-modified LDL was also partially suppressed by both EGTA and BAPTA, but their effects on the up-regulation of receptor activity by cell-modified LDL were not additive.
The enhancement of SMC SR-A expression by calcium ionophore appeared to be mediated by effects at both the transcriptional and posttranscriptional levels. Although ionophore did not augment AP-1 binding on electrophoretic mobility shift assays (data not shown), it did increase phosphorylation of c-Jun (Fig. 8), indicating more sustained activation of JNK, consistent with the promotion of AP-1/c-Jun-dependent transcription. Interestingly, the ionophore may promote alternative splicing or RNA editing that favors the SR-AII mRNA transcript. Reverse transcription-PCR of mRNA isolated after treatment with H/V or cell-modified LDL showed a synergistic increase in SR-AII mRNA with the addition of ionophore (Fig.  9). A slight but reproducible increase in SR-AI mRNA was observed under the same conditions that strongly up-regulated SR-AII mRNA; however, ionophore did not further increase SR-AI mRNA (Fig. 9).

COX-2 Up-regulation Correlates with Enhanced SR-A Expression in the Presence of Calcium
Ionophore-Because the COX-2 and SR-A gene promoters have similar transcription factor binding sites (22,33), we hypothesized that COX-2 expression would be regulated by factors that increased SR-A expression in SMC. Indeed, both H/V and cell-modified LDL increased COX-2 mRNA levels, and the increase after either treatment was further enhanced by calcium ionophore (Fig. 9).
Next, to determine whether the increase in COX-2 activity is important for the up-regulation of SMC SR-A activity, we evaluated the effects of two COX-2 inhibitors on the uptake of DiI-labeled AcLDL by SMC treated with H/V plus ionophore (Fig. 10A). With H/V plus A23,187 alone, the mean fluorescence intensity of the cells was 292, indicative of a high level of uptake of DiI-labeled AcLDL. Both NS-398, a specific inhibitor of COX-2 activity, and Resveratrol, a phenolic antioxidant that inhibits COX-2 at the transcriptional and posttranscriptional FIG. 5. Effects of cell-modified LDL on AP-1 binding and JNK activity. A, electrophoretic mobility shift assays showed a time-dependent increase in AP-1 binding in nuclear extracts from human SMC incubated with THP-1 cell-modified LDL. A 32 P-labeled AP-1 consensus binding site oligonucleotide was used as probe. For supershift analyses, c-Jun or c-Fos antibodies were added after the addition of probes. Competitions were performed with unlabeled oligonucleotide of either the AP-1 sequence or an unrelated AP-2 sequence. B, autoradiograms showed increased JNK activity within the SMC lysates after incubation with conditioned medium from THP-1 cells. The glutathione S-transferase-c-Jun fusion protein was phosphorylated with [␥-32 P]ATP that was included in the kinase reaction buffer.

FIG. 6. Effect of cell-modified LDL on C/EBP binding.
A, electrophoretic mobility shift assay showing increased C/EBP binding in nuclear proteins (5 g) isolated from human SMC treated with THP-1 cell-modified LDL for the indicated times. The SR-A promoter sequence from Ϫ44 to Ϫ21 bp, a C/EBP binding element, was used as probe. B, supershift assay showing C/EBP␤ binding activity in nuclear extracts isolated after both short (0.5-h) and long (18-h) incubations with cell-modified LDL. The binding activities were supershifted with C/EBP␤ antibody but not with antibody to GADD 153. Specificity of binding was demonstrated by complete competition with unlabeled oligonucleotide corresponding to the C/EBP wild-type sequence but not the mutant sequence. levels (33), essentially blocked the ionophore-induced increase in SR-A activity, shifting the mean relative fluorescence intensity to 16 and 19, respectively. The mean basal fluorescence was 17. Both inhibitors also substantially inhibited the ionophore-induced increase in SR-A activity in SMC treated with cell-modified LDL (Fig. 10B), reducing the mean fluorescence intensity from 123 for the cells treated with cell-modified LDL plus A23,187 to 31 (NS-398) and 53 (Resveratrol).
Although treatment with ionophore alone significantly increased COX-2 mRNA levels ( Fig. 9), treatment of SMC with ionophore alone did not increase SR-A mRNA or activity. These data suggest either that increased COX-2 activity is not sufficient for up-regulation of SMC SR-A or that increased COX-2 mRNA does not necessarily correlate with COX activity. We FIG. 7. Effect of calcium flux on SMC SR-A activity. A, human SMC were incubated overnight at 37°C with H/V or cell-modified LDL (CM LDL) (50 g/ml), with and without the calcium ionophore A23,187 (0.5 M). After treatment, cells were washed, incubated at 37°C for 15-18 h with medium containing DiI-labeled AcLDL (5 g/ml), and prepared for fluorescence-activated cell sorting analysis as described previously (11). After treatment, the cells were washed and then incubated at 37°C for 15-18 h with medium containing DiI-labeled AcLDL (5 g/ml), and the uptake of DiI-labeled AcLDL was quantitated by fluorescence-activated cell sorting analysis. Data from a representative experiment are shown. This experiment was repeated with similar results. therefore analyzed the conditioned medium for PGE 2, a specific product of COX-mediated AA metabolism, to determine if the increased COX-2 mRNA levels were associated with the production of bioactive COX-2 metabolites (Table I). PGE 2 production did not always correlate with increased COX-2 mRNA levels. Although COX-2 mRNA levels were similar in the presence of ionophore alone and ionophore with H/V (Fig. 9), PGE 2 levels were 9.2-fold higher in the latter case (Table I). Increased COX-2 activity as measured by PGE 2 production correlated with increased SR-A expression. In SMC treated with H/V and ionophore, both NS-398 and Resveratrol significantly inhibited the ionophore-induced increases in PGE 2 levels and, as described above, the enhancement of SR-A up-regulation.
Similar results were obtained with cell-modified LDL ( Table  I). The conditioned medium containing cell-modified LDL had a PGE 2 level of 2560 pg/ml, reflecting secretion into the medium by the THP-1 cells. However, when SMC were incubated in this conditioned medium, the PGE 2 level increased 3.5-fold (Table  I). Coincubation with Resveratrol or NS-398 abolished this increase. Resveratrol and NS-398 also blocked the calcium ionophore-stimulated production of PGE 2 . Therefore, under conditions in which ionophore increases SR-A activity, both COX-2 mRNA and PGE 2 levels in the medium were considerably increased. This finding suggests that the metabolic products of AA metabolism contribute to increased SR-A expression. Inhibition of SMC COX-2 activity and PGE 2 production correlated with decreased SR-A activity.
While PGE 2 is a readily identifiable marker for COX-2 activity, it is not likely to be the sole eicosanoid responsible for increased COX-2-dependent SR-A expression. Purified PGE 2 added directly to SMC did not induce SR-A activity (data not shown). Because PGE 2 is only one of a large family of eicosanoids downstream of COX-catalyzed AA metabolism, its identification in the medium implies the presence of a full complement of prostaglandins. These findings suggest that H/V treatment triggers enzymatic AA metabolism in SMC and that ionophore significantly enhances this process. Much higher levels of AA metabolites generated by COX-2-mediated pathways can be identified after SMC treatment with cell-modified LDL. The addition of ionophore minimally increased the levels of AA metabolites beyond those achieved with cell-modified LDL alone (Table I). The full biological consequences of AA metabolites are not known, but the inhibition data presented in Table I suggest that the eicosanoids contribute to SMC SR-A expression.
Cell-modified LDL, but Not Native LDL, Induce Adhesion, Monocyte-to-macrophage Differentiation, and SR-A Expression in THP-1 Cells-Next, we determined whether cell-modified or chemically oxidized LDL, like H/V treatment (22), would induced monocyte differentiation and up-regulate SR-A expression. Treatment of nonadherent THP-1 cells with cell-modified LDL (50 g/ml) induced both cell adherence and SR-A gene expression (Fig. 11). Treatment with copper-oxidized LDL resulted in the adherence of fewer cells but had a qualitatively similar effect on SR-A activity; native LDL had no effect. Interestingly, in THP-1 cells, treatment with H/V or cell-modified LDL markedly increased both SR-AI and SR-AII mRNA levels. Unlike the observation in SMC, calcium ionophore did not appreciably enhance these high levels of induction (data not shown). DISCUSSION This study shows that coincubation with macrophages or with macrophage-conditioned medium increases SMC SR-A activity. Interestingly, this effect was mediated by cell-modified LDL, not by macrophage-secreted cytokines or growth factors. Chemically oxidized LDL also increased SMC SR-A activity. The SR-A gene expression mediated by cell-modified LDL resulted from increased levels of the redox-sensitive transcription factors AP-1/c-Jun and C/EBP␤ and from increased JNK activity. These transcription factors are important for SMC SR-A up-regulation by phorbol esters and by ROS (22). In addition, cell-modified LDL, but not native LDL, induced THP-1 cell adhesion and SR-A expression. The increase in SR-A activity in SMC was enhanced by calcium ionophore and was associated with increased COX-2 activity. These findings demonstrate that oxidized lipoproteins in the arterial wall can up-regulate the scavenger receptors responsible for their clear-  11. Effect of treatment with cell-modified or Cu 2؉ -oxidized LDL on THP-1 cells. THP-1 cells in suspension were incubated for 24 h at 37°C with the indicated LDL treatments (all at 100 g/ml) and washed, and the adherent cells were incubated for 15-18 h with DiI-labeled AcLDL (5 g/ml). Both phase (left panels) and fluorescence (right panels) photomicrographs are shown. The modified LDL increased adhesion and SR-A activity; native LDL had no effect. ance by both SMC and macrophages. Increased scavenger receptor activity is thought to contribute to cholesterol ester accumulation and foam cell formation.
The regulation of redox-sensitive transcription factors by cell-modified LDL suggests that oxidatively modified lipoproteins are a source of intracellular oxidative stress. This possibility is consistent with the increased levels of intracellular lipid peroxidation products and ROS in endothelial cells treated with oxidized LDL (34). Multiple lines of defense against oxidative stress in tissues and cells are overcome in pathological conditions such as atherosclerosis (35). The effect of oxidized LDL on redox balance in the vascular space depends upon the type and extent of oxidation (36). Minimally modified LDL induce the expression of genes encoding macrophage colony-stimulating factor (37), monocyte chemoattractant protein 1 (38), and tissue factor (39), all of which, like the SR-A gene (22), are induced by redox-sensitive transcription factors (40 -42). While mounting evidence suggests that oxidants play a critical role in the regulation of gene expression, the pathways for redox-regulated signal transduction remain largely speculative (16). Oxidized LDL up-regulate the class B scavenger receptor CD36 and the SR-AI and SR-AII in macrophages (43,44). In macrophages, CD36 expression, but not SR-A expression, results from peroxylipid activation of peroxisome proliferator-activated receptor ␥-dependent transcription (45). No treatment in the present study induced CD36 in SMC (data not shown). The transcriptional pathways leading to SR-A induction in either macrophages or SMC by oxidized LDL have not been defined. However, oxidized LDL induce C/EBP␤ in monocyte/macrophages (46) and AP-1 in human arterial SMC (47). Both of these transcription factors are important for SMC SR-A induction by treatments that induce intracellular oxidative stress (11,22) and, as shown here, are increased by moderate cell-mediated oxidation of LDL. AP-1 up-regulation by oxidized LDL in SMC has been attributed to the lysophosphatidylcholine content of the oxidized lipoprotein (47). Lysophosphatidylcholine also induces intracellular calcium flux (48). Extensive oxidation of LDL leads to cytotoxicity and apoptosis (36), cellular processes also associated with calcium flux (49).
Oxidants stimulate Ca 2ϩ signaling by increasing cytosolic Ca 2ϩ concentration (17,18). The cellular oxidative stress generated by oxidized LDL has been associated with a rise in free cytosolic calcium (50). While the exact source of calcium release has yet to be defined, enhanced Ca 2ϩ transport through Ca 2ϩ channels (51), inhibition of Ca 2ϩ pumps (52), and Ca 2ϩ release from intracellular stores (17) have all been described in the presence of oxidants. Our current findings demonstrate that SMC SR-A gene activation is strikingly dependent on calcium flux. SR-A gene expression was effectively inhibited by the nonspecific calcium channel blocker nickel chloride and enhanced by the calcium ionophore A23,187. Specific calcium L and T type channel blockers, however, did not block SR-A up-regulation as NiCl 2 did. NiCl 2 is a potent, but nonspecific, calcium channel blocker whose mechanism of action is not completely understood. Besides blocking the influx of calcium into cells, NiCl 2 inhibits intracellular calcium mobilization and phospholipase C activation (53). The virtually complete suppression of SR-A up-regulation by NiCl 2 may result from a combination of these effects.
The importance of calcium mobilization for the redox regulation of SR-A gene expression is underscored by the ability of calcium chelators to suppress SR-A induction by H/V plus A23,187 (Fig. 7). It is not evident why SR-A up-regulation by cell-modified LDL was only partially inhibited by the same combination of intra-and extracellular calcium chelators. Possibly, the THP1-generated eicosanoids already present in me-dium containing THP1-cell-modified LDL (see Table I) can promote intracellular oxidative stress without additional calcium mobilization. The mechanism of further calcium mobilization by cell-oxidized LDL has not been established. It is possible that bioactive lipid peroxides affect cell membranes, physically disrupting ionic homeostasis (54,55). Because oxidized LDL induce a significant calcium flux, calcium ionophore augments SR-A expression to a lesser extent than was observed after treatment with H/V.
Membrane lipid peroxidation is associated with phospholipid hydrolysis by Ca 2ϩ -dependent phospholipase A 2 (PLA 2 ) (56). An increased net negative charge of membranes increases the binding affinity for Ca 2ϩ (57), serving to catalyze PLA 2 (58). In macrophages, a sustained increase in intracellular calcium induced by A23,187 is sufficient to activate PLA 2 and release AA from phospholipids (59). Reasoning that this may also be true in SMC induced to take on macrophage-like properties, such as SR-A gene expression, we evaluated certain bioactive products of AA metabolism. AA is specifically metabolized by COX-1 and COX-2. Normally, the majority of COX expressed within vessels is in the constitutive form, COX-1, found mainly in the endothelial layer, with much lower levels being present in the underlying smooth muscle (60). However, data from animal models have suggested that when the endothelium is compromised or vessels are damaged, the inducible isoform, COX-2, can be expressed in vascular smooth muscle (61). Furthermore, the COX-2 promoter, like the SR-A promoter, contains a critical redox AP-1/c-Jun binding site as well as an upstream CEBP/␤ binding element (33). And the COX-2 gene, like the SR-A gene, is markedly induced after stimulation with cytokines, growth factors, or tumor promoters (62)(63)(64).
We found that factors that increase SR-A activity up-regulated COX-2 mRNA and that COX-2 activity correlated with increased SR-A expression. Furthermore, COX-2 inhibitors blocked the SR-A up-regulation. Calcium ionophore may at least in part mediate the enhancement of SR-A gene expression by providing AA substrate for COX-2. COX-2 induction alone does not greatly increase prostanoid production in vivo. A second, AA-liberating stimulus is also required (65). A concomitant increase in calcium-dependent PLA 2 activity and mobilization of AA is presumed to supply the COX-2 enzymatic system with substrate.
The extracellular group II PLA 2 (sPLA 2 ) has been implicated in numerous inflammatory conditions, and transgenic mice overexpressing sPLA 2 exhibit dramatically increased atherosclerosis on a high fat diet (66). This increase has been attributed to the finding that polyunsaturated free fatty acids, which are liberated by sPLA 2 , increase the formation of bioactive phospholipids in oxidized LDL (67). PLA 2 has also been implicated in the up-regulation of SR-A expression in cultured aortic SMC (68). Interestingly, although a sustained increase in calcium is sufficient to induce AA release (59), we found that ionophore alone is not sufficient to induce significant SR-A expression. The ability of calcium ionophore to up-regulate COX-2 expression in SMC indicates that the presence of COX-2 and AA together is also insufficient for SR-A induction. Ionophore treatment alone did not affect AP-1/c-Jun binding (data not shown) or JNK activity (Fig. 5B). Therefore, concomitant up-regulation of AP-1/c-Jun and PLA 2 -mediated AA release and COX-2 induction appear to be necessary for maximal upregulation of SMC SR-A activity. The mechanism for this synergy remains under investigation. It is tempting to speculate that COX-2-mediated metabolism of AA contributes to intracellular oxidative stress by the generation of vasoactive eicosanoids. Indeed, cytomegalovirus-induced generation of oxidative stress in SMC has been attributed to COX-2 activity (69). Cytomegaloviral infection of SMC has also been shown to lead to SR-A expression (70). A schematic representation of the interrelated pathways proposed for the regulation of SR-A expression in SMC is shown in Fig. 12.
Our results suggest that antioxidants or aspirin and other nonsteroidal anti-inflammatory therapies may prevent atherosclerosis at least in part by decreasing vascular oxidative stress and expression of SR-A and COX-2. The attenuation of COX-2-mediated redox-sensitive SMC gene expression has been attributed both to the direct free radical scavenging activity of aspirin and indomethacin and to their activity as COX inhibitors (69). Calcium channel blockers may exert their protective effects farther downstream by diminishing oxidant-induced calcium flux. Primary prevention of LDL oxidation, of course, would be preferable to any of these therapies. Our findings show that mild LDL oxidation triggers redox-sensitive SR-A transcriptional machinery, activating SR-A expression in both SMC and macrophages and effectively priming these cells for the clearance of fully oxidized LDL. Removal of the proinflammatory stimulus of peroxylipids from the vascular space should prevent oxidant-induced calcium flux, COX-2 induction, PLA 2 activation, AA release, and ultimately redox-sensitive SR-A expression and foam cell formation, all of which should significantly retard atherogenesis.

FIG. 12. Schematic representation of pathways proposed for the regulation of SR-A expression in SMC.
Exogenous oxidants such as oxidized LDL (OxLDL) may disrupt cell membrane integrity, inducing calcium flux and triggering intracellular oxidases. The resultant increase in the generation of ROS can inhibit protein-tyrosine phosphatase (PTP) activity, leading to sustained protein-tyrosine kinase (PTK) activity. This engages the mitogen-activated protein kinase (MAPK) cascade that is associated with phosphorylation of p38 MAPK, resulting in C/EBP␤ phosphorylation and activation; phosphorylation of JNK, resulting in AP-1/c-Jun phosphorylation and activation; and phosphorylation of cytoplasmic phospholipase A 2 (cPLA 2 ), resulting in the release of AA from phospholipid (PL) producing lysophospholipid (LPL). Membrane lipid peroxidation together with calcium flux may also activate Ca 2ϩ -dependent PLA 2 with further generation of AA. The activation of AP-1/c-Jun-and C/EBP␤-dependent gene transcription results in increased SR-A and COX-2 expression. Increased SR-A activity would contribute to the uptake of oxidized LDL and the generation of foam cells, while increased COX-2 activity could act upon the increased stores of AA substrate and generate a large family of inflammatory eicosanoids. Eicosanoid activity may further contribute to intracellular oxidative stress and, in this way, contribute to SR-A up-regulation.